An open notebook · Cancer, agents & uncertainty
Make care work.
Make time count.
Keep living.
I’m navigating rare cancer in a medical system that can be slow, fragmented, and prone to mistakes. My main job is coordinating care—and protecting doctor time for the decisions that need it. This is what I’m learning.
Written as The Bayesian Cancer Patient. Shared anonymously.
My approach
Orchestrate the system.
Protect doctor time.
The best analysis cannot help if the right doctor never sees it. I use agents and careful follow-up to get the right facts, records, and questions into the clinical conversation.
The complete working notebook- 01
Make the next decision clear.
I start with the choice, the deadline, and the missing facts. “Learn everything” is not a finishable task.
- 02
Track each handoff to completion.
A request is not receipt. Confirm which office has the report, images, or tissue, what is missing, and who owns the next step.
- 03
Put disagreements on the page.
Different pathology readings and doctor perspectives need explicit comparison. Keep the source, interpretation, and unresolved question together.
- 04
Save doctor time for judgment.
Bring a short case brief and the few questions that can change care. Let agents handle information work, and leave room for recovery and life.
Built for agents, readable by people
The whole notebook. One Markdown file.
Case context, practical lessons, research methods, reusable prompts, sources, and open questions. Give your agent this page as a starting point.
What I’m learning
What helped me make progress.
The methods I’m using, the mistakes I’m trying to avoid, and the work I give to agents.
Make the medical system work
A slow, fragmented system needs clear owners, checked handoffs, and protection for doctor time.
5 min read · Care logisticsLet trusted people in
I told three friends and three colleagues. That brought company, visits, and introductions to people with related experiences.
3 min read · People & supportUse my self-referral template
An anonymized template adapted from messages I actually sent, with a clear ask and a short case summary.
4 min read · Care logisticsCompare pathology and doctor perspectives
The actual differences in my pathology reviews, doctor biases, and how I compare recommendations.
7 min read · Care logisticsSee an analysis I actually ran
How a Claude analysis organized the chemotherapy question—and where its conclusions outran the evidence.
6 min read · Understanding evidenceLet agents do the information work
Summarize records, find gaps, and prepare questions—with a clear boundary around clinical decisions.
4 min read · Agents & organizationTrack the handoff, not the request
Reports, scan images, pathology slides, and the handoffs that can slow everything down.
4 min read · Care logisticsAsk whether the evidence fits
Check the cancer type, treatment setting, comparison, and outcome before using a result.
5 min read · Understanding evidenceGive every test a decision to inform
dMMR, inherited risk, and blood-based monitoring are related, but they answer different questions.
4 min read · Understanding evidenceKeep one current case summary
Build a one-page view of what is known, what is pending, and who owns the next step.
4 min read · Agents & organizationOne question worth bringing
“What evidence would change your recommendation?”
A simple way to find the assumptions, the missing information, and the next useful step.
Get the appointment checklistGo to the source
A small, considered reading list.
Understanding appendix cancer
A starting point for the disease and its different forms.
The ATOMIC trial
Why a promising result needs a careful discussion of case fit.
Making sense of prognosis
What group statistics can and cannot tell an individual.
Other patient voices
People worth learning from.
Different experiences and approaches. Personal accounts are context, not proof of a treatment’s effect.
Kevin Wimer
His practical, patient-first website helped shape the idea for this one.
Mike Lynn
A public voice to include in the broader patient-learning conversation.
Sid Sijbrandij
A detailed account of a complex care and research effort.